This program of research addresses psychological/health status relationships in children with insulin dependent diabetes mellitus. Psychological variables that should have a direct bearing on the child's health were selected for initial study. A causal model was developed in which patient and parent knowledge about diabetes, patient and parent attitudes toward diabetes, and parental supervision of the child were presumed to relate to adherence with the diabetes regimen, and adherence, in turn, was presumed to relate to diabetes control. Although adherence was predicted reasonably well by the model, the link between adherence and diabetes control proved problematic. In a series of cross-sectional and longitudinal investigations, few significant associations between adherence and diabetes control emerged. When significant associations did occur, they were weak in magnitude and/or highly complex, involving interactions with other variables (e.g., patient age, disease duration, diabetes control at the study's inception). The proposed application explores adherence/diabetes control linkages further using a longitudinal research design with adolescents who have had the disease for 3 years or more. The disease duration requirement was selected to eliminate variability due to endogenous insulin replacement in youngsters more recently diagnosed. The longitudinal design was selected to permit measurement of change in the model constructs selected for study. Adolescents are of particular interest because they are less adherent and are in worse diabetes control than their younger counterparts. Yet, prior causal models have failed to identify strong, consistent and significant paths between adolescents' adherence behaviors and their diabetes control. Maturational changes associated with puberty may exert a negative influence on diabetes control during this developmental period; this hypothesis will be tested in the research proposed. Once these biological changes are statistically modelled, adherence/diabetes control relationships may be more readily discerned. However, it is also possible that our prior conceptualization of adherence is faulty. To address this issue, four different conceptualizaions of adherence will be compared. Similarly, measures of diabetes control have been expanded in an effort to tap adherence/diabetes control linkages in a more sensitive manner. Finally, this application proposes to examine both main effects of adherence, pubertal status, and insulin dose and to explore possible interactions between adherence and the other model constructs, as predictors of diabetes control.